Análisis desde la perspectiva de la modelización científica

CM differs from the conventional medicine in a number of ways, including the diagnostic methods used. CM diagnostic approaches rely more on the clinician’s interpretation of the patient’s symptoms and signs rather than on laboratory tests.(Zhang et al., 2003) The process of CM diagnosis is the collection of patient information through taking a case history and an overall observation of symptoms and signs, analysis of these according to CM theories and drawing a conclusion about the disease and CM Syndrome.(Deng et al., 1984, Maciocia, 2004) As discussed previously, there are four diagnostic methods used in CM diagnosis: Inquiry, Inspection, Auscultation/Olfaction, and Palpation.

Up to now, only one study has investigated the reliability of the CM diagnostic process in a comprehensive way at both the level of data collection (using three of the four diagnostic methods, Palpation, Inspection and Auscultation) and data analysis or Syndrome differentiation (diagnosis using the Eight Guiding Principles and Zang-Fu Theory).(O'Brien et al., 2009a, O'Brien et al., 2009b) Other studies have focussed on the consistency of individual diagnostic methods such as pulse diagnosis or tongue diagnosis or on the CM Syndrome diagnosis only.

This section will summarise the current knowledge about the reliability of the four diagnostic methods, and CM Syndrome diagnosis according to the Eight Guiding Principles and Zang-Fu Theory.

6.2.1 Inspection

Only one comprehensive inter-rater reliability study has assessed the reliability of various aspects of visual inspection in a CM examination.(O'Brien et al., 2009b) In this study, diagnostic variables of inspection included spirit (shen), complexion,

126 colour around the eye, hair, tongue body and tongue coating. The researchers found various level of agreement amongst the three practitioners, from slight (0.00 < κ≤ 0.20) in the case of colour around the eye and skin texture, to fair (0.21 < κ≤ 0.40) for colour of complexion, to almost perfect agreement (0.81 < κ≤ 1.00) in the case of presence of shen. The Kappa value (κ) is a measure of the level of agreement

between observers beyond that occurring by chance. When comparing between at

least two practitioners, the level of agreement for most of diagnostic outcomes of inspection was higher. For example, the level of agreement for colour of complexion was 93% (κ= 0.85) The relatively low levels of agreement for many of the inspection variables may be due to a more subtle level of observation required, and also demonstrates the subjectivity of inspection. (O'Brien et al., 2009b)

There are some inherent difficulties in measuring reliability of inspection. Firstly, the quantification of subjective judgements made by the practitioner is not easy. For example judgement of a patient’s spirit is part of Inspection, however it is hard to quantify someone’s spirit which depends on the physical health status and may change with emotional state. Secondly, descriptions of particular observation indices in CM have not been defined clearly, leading to difficulties in quantifying or measuring such items. For example, in observing the colour of the complexion, five colours (green, red, yellow, white and black) are typically described in CM texts (Deng et al., 1984), and attempts may be made to describe these further in terms of depth of colour (eg. light, dark and deep). However, the eye is able to detect dimensions such as hue, saturation and brightness. Describing the complexion in terms of five colours is limited. The inherent subjectivity of inspection may lead to low inter-rater reliability among CM practitioners.

In China, some medical instruments and computerised image analysis systems have been developed and introduced into analysis of CM tongue diagnosis.(Chen and Zhang, 2008, Wang et al., 2005, Wei et al., 2002, Weng and Huang, 2001) Most of them adopt the red/green/blue (RGB) colour model in which red, green, and blue light are added together in various combinations to reproduce a broad array of colours. Several other models have been applied in tongue research.(Gong et al., 2005, Wei and Li, 1995, Zhang et al., 2005b, Zhang et al., 2004d, Xu et al., 1993, Zhang et al., 2005d), including the ‘Lab’ model (a colour-opponent space with three dimensions,

127 dimension L for lightness and dimensions a and b designating the colour-opponent dimensions), the YUV model (Y designating the luminosity or brightness component and U and V the chromaticity or colour components), and the HSL (hue/saturation/lightness) model which attempts to describe perceptual colour relationships more accurately than the RGB and the Hue model. These models are attempts to quantify characteristics of the tongue body and tongue coating colour. Although studies have found that tongue colour identification systems can reflect the characteristics of tongue colour and record similar judgements as those of CM practitioners, there are several factors which could contribute to errors in measurement, such as the shape of tongue, the structure of curved surface and the sampling area.(Wang et al., 2005, Wei et al., 2002, Weng and Huang, 2001)

There are very few studies of the reliability of tongue inspection in English language publications.(Kim et al., 2008, O'Brien et al., 2009b, Rupp, 1998) Two of the studies utilised slides and one utilised real patients.(O'Brien et al., 2009b) One study of the inter-reliability and intra-reliability of CM tongue inspection was conducted among thirty CM practitioners, most of whom were trained in Australia with a minimum of three years of CM education and a minimum of six months independent clinical experience.(Kim et al., 2008) The practitioners completed two questionnaires based on standard colour slides and test tongue slides. The standard colour slides (five colours: dark red, pale pink, pink, red, and purple) were scanned from a standard colour chart while ten tongue photographs were selected according to different diagnostic characteristics. The findings indicated poor reliability for CM tongue inspection. The study also suggested that the low levels of inter- and intra-practitioner agreements may be due to ambiguous operational definitions of tongue colour characteristics in CM, and were not related to duration of clinical practice. There are limitations of using slides of tongues compared with real tongue presentations since qualities such as moisture may not be readily detectable in slides. The actual tongue diagnosis (like other CM diagnostic techniques) is also dependent on the experience of the CM practitioner.

O’Brien and colleagues conducted an inter-rater reliability of CM diagnosis that included tongue diagnosis as a substudy of a clinical trial investigating the efficacy of Chinese herbal medicine in the treatment for hypercholesterolemia. They reported a

128 ‘slight’ level of agreement (0.00 < κ< 0.20) in terms of tongue body characteristics (size, colour, presence of teeth marks and papillae) and a ‘fair’ level of agreement (0.21 < κ< 0.40) in terms of tongue coating characteristics (including quality, colour and thickness) amongst three CM practitioners (all of whom had at least five years experience).(O'Brien et al., 2009b). This suggests that tongue diagnosis is quite subjective and that some aspects are relatively reliable, and others not so reliable. 6.2.2 Auscultation/Olfaction

In CM clinical practice, not much significance is attached to Auscultation since the hearing ability of practitioners varies from person to person. In ancient China, there was no effective tool to capture sound information, thus as a diagnostic tool it has not developed particularly. In contrast, in conventional medicine, auscultation has become an effective way to distinguish pathological changes in the chest since the invention of stethoscope.

In CM, the intensity of voice is used to identify ‘deficiency’ or ‘full’ (‘excess’) Syndromes and the relevant zang-fu organs which may be involved in the Syndrome. For example, a very loud, barking cough typically indicates a ‘full’ Syndrome. In addition, according to CM theory, sounds also correspond to the Five Element Theory. For example, shouting is associated with Wood, laughing with Fire, singing with Earth, crying with Metal and groaning with Water.(Maciocia, 2005) But this theory has not been verified using scientific research methods.

Some research has attempted to quantify and interpret the voice using a spectrogram.(Mo et al., 1998b) One study (Mo et al., 1998a) demonstrated there were significant differences in vowel pronunciation of Mandarin words and cough sounds measured by a sound spectrogram amongst a group of healthy people and three groups of patients with cough (Deficiency of Lung Qi group, Deficiency of Lung Yin group and Full Syndrome group). The author argued that CM auscultation can be objectified by measuring the sound spectrum values of the voice including harmonic waves, crest waves, amplitude, formant, noise, radical frequency and duration of the crest wave.(Mo et al., 1998a)

Only one study published in English has assessed the inter-rater reliability of auscultation variables in a CM examination. They found ‘almost perfect’ agreement

129 for voice strength and breath sounds between at least two CM practitioners.(O'Brien et al., 2009b) However, auscultation also includes eructation and other sounds such as groaning and crying which have not been studied with respect to reliability.

6.2.3 Inquiry

In conventional medicine, the consultation between the doctor and patient is the first and essential procedure in diagnosis. In CM, this is even more important since there is less reliance on objective medical diagnostic tests (as are used in orthodox medicine).(Lu et al., 2005) As the majority of information acquired from the Inquiry component of the CM examination are subjective symptoms, it is common to find inconsistency in diagnostic assessments among different doctors when subjective judgments are involved even in conventional medicine.(Moyer et al., 2000, Santucci et al., 2000)

Subjective symptoms are usually regarded as the subjective reaction to pathological changes in the human body and may also be used as a means of subjective assessment by the patient of the effectiveness of treatment. For example, low back pain or perspiration won’t influence the diagnosis or categorisation of a hepatitis patient in conventional medicine; any change of defecation in the hepatitis patient after treatment won’t determine the treatment strategy either. However, the above- mentioned information is important information in CM Inquiry and may result in a different Syndrome diagnosis and consequently different treatment plan. Zhang and colleagues argued that inquiry as a diagnostic method is not a valid instrument, because of the low level agreement of diagnosis amongst three practitioners in the diagnosis for rheumatoid arthritis patients.(Zhang et al., 2005a, Zhang et al., 2004b) Research has found the use of relatively objective questionnaires to collect clinical data instead of simply recording notes during the taking of a case history could improve the agreement in CM diagnosis among CM practitioners for rheumatoid arthritis.(Zhang et al., 2008a) Another validation study also suggested a questionnaire designed to evaluate Yin deficiency Syndrome can help make CM diagnosis more objective and reliable in the diagnosis of that particular Syndrome.(Lee et al., 2007) In addition, another study recommended that multiple consultations could improve the agreement rates amongst different CM practitioners- repeated consultations and modification of treatment plans are common practices in CM and multiple

130 consultations can help in formulating the most appropriate diagnosis, treatment principle and regime.(Sung et al., 2004)

6.2.4 Palpation

Palpation in CM includes palpation of body parts as necessary, and taking the radial pulse- pulse diagnosis. Pulse diagnosis is a unique diagnostic technique of CM involving palpation of the pulse located at the radial artery on both wrists, assessing the characteristics of the pulse and interpreting these in relation to the functional and physiological state of the internal organs. In spite of its crucial role in the diagnostic process in CM, it is difficult to objectify and standardize pulse diagnosis. Some attempts have been made in China to try to objectify pulse diagnosis through development of pulse measuring apparatus.(Cai et al., 2007, Yang et al., 2000) However, some have argued that Zang-Fu Organ Theory and the principles of Syndrome differentiation have been ignored in the design of these machines, which are simply pulse tracing devices based on anatomy and physiology as described in conventional medicine.(Tian and Lu, 2008) In addition, the concept of Qi should be considered in the development of pulse measuring devices and used in research. In CM the pulse may change according to several factors including seasonal changes, diurnal variation, emotional state, physical health and even body fat distribution. That is, such changes are typically reflected in the pulse. Therefore, pulse diagnosis cannot be simplified as merely detecting the indexes of hydrodynamics.(Li, 2003b)

A study by Craddock (Craddock, 1997) found low levels of inter-rater and intra-rater reliability when increased complexity of pulse qualities were measured. Another study of the reliability of pulse diagnosis which standardised the pulse-taking procedures and definitions of the characteristics of each pulse showed a relatively high level of agreement between two independent pulse assessors in the first collection session in 66 subjects (81% agreement) and in the replicated session in 30 subjects two months later (80% agreement).(King et al., 2002) O’Brien and colleagues found a ‘fair’ level of agreement (κ= 0.29) among three practitioners in term of pulse force. (O'Brien et al., 2009b) However, variable findings have been reported in other studies. A review of published and unpublished reliability studies of pulse diagnosis (O'Brien and Birch, 2009) indicates that the reliability varies from a low level to a very good agreement. However caution should be taken in

131 interpretation of those findings because different approaches have been used to collect, analyse and report the data. For example, in one study CM students were used as raters; a low level of agreement was found at the beginning of formal pulse diagnosis teaching classes (week 1), at the conclusion of pulse teaching (week 14), and one year later.(Walsh et al., 2001) It could be argued that lack of clinical experience may reduce the possibility of agreement.

Current research indicates that unambiguous definitions of pulse characteristics is critical in pulse research.(King et al., 2006, O'Brien and Birch, 2009) Standardising the pressure applied (that is, measurement of finger strength) in pulse detection is the most difficult part in pulse research. (Tian and Lu, 2008) However, the integration of an ultrasound technique into a pulse detection device that could generate multi- dimensional tomograms of the pulse could be a new direction of pulse research.(Liu, 2003)